Elemental soil plot tracing method based on nonparametric test

Abstract

The invention discloses an elemental soil plot traceability method based on nonparametric inspection, and relates to the technical field of plot traceability methods.The elemental soil plot traceability method comprises the following steps that A1, first-time soil collection is conducted, and geochemical elements in a first-time soil collection sample are detected; a2, carrying out secondary soil collection, and detecting geochemical elements in a secondary soil collection sample; a3, the first-stage data and the second-stage data are analyzed and processed; a4, comparing and analyzing the first-stage data and the second-stage data, and constructing a traceability model; a5, geochemical elements of the soil sample needing to be traced are detected, data obtained through detection are substituted into the traceability model to be compared, and the collection site of the soil sample is determined. According to the scheme, the soil plot can be positioned.

Description

Technical Field

[0001] The present invention relates to the technical field of land plot tracing methods, and in particular to a land plot tracing method for locating land plots using geochemical elements. Background Art

[0002] With the high-quality development of geographical indications, the scale of the geographical indication industry continues to expand. The promotion of geological indications and characteristic industries is inseparable from the identification of geographical characteristics by soil geochemical information. The processing of geographical soil information data will be very numerous, which will lead to the soil data being relatively chaotic, making it impossible to effectively know the geographical source of the geochemical elements of the soil. Based on the key technology of soil geochemical element plot positioning data analysis, a technical method for soil plot positioning is needed to solve the above-mentioned problems. Summary of the invention

[0003] The purpose of the present invention is to provide a method for tracing the origin of a land block by locating geochemical elements in view of the deficiencies in the prior art.

[0004] A method for tracing the origin of elemental soil plots based on nonparametric testing, comprising the following steps:

[0005] A1: Carry out the first soil collection and test the geochemical elements in the first soil collection sample;

[0006] A2: Conduct a second soil collection and test the geochemical elements in the second soil collection sample at the same location;

[0007] A3: Process the data of the first phase and the second phase;

[0008] A4: Compare and analyze the data of Phase I and Phase II and build a traceability model;

[0009] A5: Test the geochemical elements of the soil samples that need to be traced, bring the test data into the traceability model for statistical analysis, and determine the location of the element numbers belonging to the collection site of the soil sample;

[0010] The analysis and processing of the first-phase data and the second-phase data include data processing by sample data normalization method, sample element runs test, sample element descriptive statistical analysis, sample element Kendall test, first-phase and second-phase sample element paired sample correlation analysis, T test, and plot and element correlation test.

[0011] In some embodiments, the formula for data processing of the sample data normalization method is as follows:

[0012]

[0013] in:

[0014] X is the original data value,

[0015] Xmin is the minimum value in the data set.

[0016] Xmax is the maximum value in the data set,

[0017] Xnorm is the normalized data value.

[0018] In some embodiments, the calculation formula for the sample element runs test is as follows:

[0019] R = total number of runs;

[0020] n1 = number of observations in the first category;

[0021] n2 = number of observations in the second category;

[0022] The expected number of runs E(R) and variance V(R) are:

[0023]

[0024]

[0025] Among them, n1+n2=n is the total sample size.

[0026] In some embodiments, the sample element descriptive statistical analysis can perform statistics on the generated data, and provide a preliminary understanding of the overall geochemical data by summarizing and describing the basic characteristics of the data. Through descriptive statistics, errors, outliers or unreasonable records in the data can be discovered, thereby performing data cleaning. Descriptive statistics can provide preliminary evidence and support for subsequent hypothesis testing.

[0027] In some embodiments, the sample element Kendall test is used to evaluate whether two random variables are statistically consistent. The following are the steps for calculating the Kendall W coefficient based on two variables X and Y:

[0028]

[0029] Where n is the number of subjects, k is the number of raters, is the average rating of the ith object, It is the average of the average scores of all objects. The value range of Kendall's W coefficient is between 0 and 1. The closer the value is to 1, the higher the consistency between the scorers, and the closer the value is to 0, the lower the consistency. It is compared with the standard normal distribution to determine the fluctuation difference of element data and prove the significant characteristics of inconsistency, which can provide data support for subsequent cluster analysis.

[0030] In some embodiments, the first-stage and second-stage sample element paired sample T test comprises the following steps:

[0031] F1: Calculate paired differences: For each pair of observations, calculate their difference;

[0032] F2: Calculate the mean of the differences: Calculate the average of all paired differences;

[0033] F3: Calculate the standard deviation of the difference: Calculate the standard deviation of all paired differences;

[0034] F4: The formula for calculating the T statistic is as follows:

[0035]

[0036] in:

[0037] is the sample mean of the differences;

[0038] μ0 is the population mean of the differences;

[0039] SD is the standard deviation of the difference;

[0040] n is the number of pairs;

[0041] F5: Determine the degrees of freedom: The degrees of freedom (df) is the number of pairs minus 1(n-1).

[0042] In some embodiments, the method for verifying the correlation between a land parcel and an element includes:

[0043] B1: Data collection, collect two sets of data;

[0044] B2: Calculate the paired difference. For each pair of data, calculate their difference;

[0045] B3: Test the distribution of differences and whether these differences obey normal distribution;

[0046] B4: Calculate statistics, including the mean value Dˉ and standard deviation SD of the difference. The main formula is:

[0047]

[0048] Where n is the number of paired samples;

[0049] B5: Determine the degrees of freedom, the degrees of freedom (df) is the number of paired samples minus 1(n-1);

[0050] B6: Conduct hypothesis test, set the null hypothesis (H0): there is no significant difference between the mean values ​​of the two samples, set the alternative hypothesis (H1): there is a significant difference between the mean values ​​of the two samples, and find the corresponding p value in the T distribution table based on the T statistic and degrees of freedom;

[0051] B7: Draw a conclusion: If the p-value is less than the significance level, reject the null hypothesis and conclude that there is a significant difference between the mean values ​​of the two samples. If the p-value is greater than the significance level, fail to reject the null hypothesis and conclude that there is insufficient evidence to show that there is a significant difference between the mean values ​​of the two samples.

[0052] In some embodiments, the sample element second-order clustering analysis includes a pre-clustering stage and a model clustering stage.

[0053] In the pre-clustering stage, a distance-based algorithm is used to create smaller clusters in order to reduce the amount of subsequent calculations. In the model clustering stage, the intra-class covariance matrix is ​​used. The intra-class covariance matrix of each cluster is used to describe the variability within the cluster. The element clustering uses the model selection criteria to determine the final number of clusters and cluster members. The Bayesian Information Criterion (BIC) is used to evaluate the goodness of fit of different clustering models. The BIC formula is as follows:

[0054] BIC=-2log(L)+klog(n)

[0055] in,

[0056] L is the maximum likelihood estimate of the model,

[0057] k is the number of free parameters in the model,

[0058] n is the sample size.

[0059] In some embodiments, the method for constructing the sample lateral positioning inspection model includes:

[0060] C1: Data collection, collect two sets of data, each set of data contains the same number of observations, and these observations are paired;

[0061] C2: Calculate the paired difference. For each pair of observations, calculate their difference.

[0062] C3: Calculate the sign rank and absolute rank of the difference. For each difference, calculate its sign rank and absolute rank.

[0063] C4: Calculate statistics, calculate the rank sum (RS), including the positive rank sum (R+) and the negative rank sum (R-). Use the following formula to calculate the T statistic:

[0064]

[0065] Where n is the number of pairs;

[0066] C5: Determine the degrees of freedom, the degrees of freedom (df) is the number of pairs minus 1(n-1);

[0067] C6: Conduct hypothesis test, null hypothesis (H0): there is no significant difference in the mean values ​​of the two samples, alternative hypothesis (H1): there is a significant difference in the mean values ​​of the two samples, find the corresponding p value in the Wilcoxon distribution table based on the T statistic and degrees of freedom;

[0068] C7: Draw a conclusion. If the p-value is less than the significance level, reject the null hypothesis and conclude that there is a significant difference between the mean values ​​of the two samples. If the p-value is greater than the significance level, fail to reject the null hypothesis and conclude that there is insufficient evidence to show that there is a significant difference between the mean values ​​of the two samples.

[0069] C8: Perform geochemical analysis based on the detection characteristics.

[0070] In some embodiments, the sample element longitudinal plot test model is established to test whether there is a significant difference in the medians of three or more independent samples.

[0071]

[0072] in,

[0073] N is the total number of observations,

[0074] k is the number of samples,

[0075] ni is the number of observations in the ith sample,

[0076] Ri is the rank sum of the i-th sample,

[0077] According to the value of the H statistic and the corresponding distribution table, determine whether to reject the null hypothesis, that is, the number of locations of all samples is equal. If the p value is less than the pre-set significance level, the null hypothesis is rejected, and it is considered that the number of locations of at least one sample is different from that of other samples.

[0078] The advantages of the present invention compared with the prior art are:

[0079] In this scheme, the method of converting geochemical elements into statistical analysis is used to trace the origin of soil plots. Geochemical elements become digital elements and are normalized to eliminate the order of magnitude differences between element contents, which can make the program converge faster when running. Element data are tested for runs to determine whether the element content conforms to random distribution or has a certain pattern or trend. Element runs test is performed to determine the hypothesized characteristics of the randomness of element values, including descriptive statistical analysis, Kendall test, paired sample T test, interactive analysis of plots and elements, and other studies. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] Figure 1 A flowchart structure diagram of the method for tracing the origin of geochemical elements;

[0081] Figure 2 The main technical flow chart of the method for tracing the origin of geochemical elements;

[0082] Figure 3 It is a schematic diagram of data processing of sample data normalization method;

[0083] Figure 4 It is the significance value diagram of sample element runs test;

[0084] Figure 5 It is the second-order cluster analysis diagram of elements;

[0085] Figure 6 Locate elements in the overall element range map when tracing the soil.

[0086] Figure 7 To locate the elements in the overall plot when tracing the soil. DETAILED DESCRIPTION

[0087] Combined with Figure 1-7 As shown in the figure, the method of converting geochemical elements into statistical analysis is used to trace the land source of soil. Geochemical elements become digital elements and are normalized to eliminate the order of magnitude difference between element contents, which can make the program converge faster when running. The element data is tested for runs to determine whether the element content conforms to the random distribution or has a certain pattern or trend. The element runs test is performed to determine the hypothetical characteristics of the randomness of the element value, including descriptive statistical analysis, Kendall test, paired sample T test, interactive analysis of land plots and elements, etc. The main technical processes are as follows Figure 2 As shown,

[0088] A method for tracing the origin of elemental soil plots based on nonparametric testing, comprising the following steps:

[0089] A1: Carry out the first soil collection and test the geochemical elements in the first soil collection sample;

[0090] A2: Conduct a second soil collection and test the geochemical elements in the second soil collection sample;

[0091] A3: Analyze and process the data of Phase I and Phase II;

[0092] A4: Compare and analyze the data of Phase I and Phase II and build a traceability model;

[0093] A5: Test the geochemical elements of the soil samples that need to be traced, bring the test data into the traceability model for comparison, and determine the collection site of the soil samples;

[0094] In steps A1 and A2, the soil collection method can be based on the standard of Technical Guidelines for Soil Sampling of Soil Quality (GBT36197-2018), which provides technical guidance on the collection and storage of soil samples. According to the characteristics of the plot and the purpose of the investigation, appropriate sampling points should be selected. It is usually necessary to consider factors such as the size, shape, soil type and known pollution distribution of the plot. Surface samples should be collected away from the fertilization period. At each sampling point, soil is usually collected from multiple sub-points (such as diagonals or grid points) to obtain a more uniform sample. If the plot is large, random or systematic sampling methods should be used to select sampling points. The distribution of samples can represent the main environmental types within the sampling grid.

[0095] Place the collected soil samples in clean, labeled sample bags. If necessary, perform preliminary sample processing on site, such as removing stones, plant roots, etc. Clearly mark the sampling point number, sampling date, sampling depth and other information on each sample bag.

[0096] The analysis and processing of the first-phase data and the second-phase data in step A3 include data processing by sample data normalization method, sample element runs test, sample element descriptive statistical analysis, sample element Kendall test, first-phase and second-phase sample element paired sample T test, and plot and element correlation test;

[0097] like Figure 3 As shown, the formula for data processing of the sample data normalization method is as follows:

[0098]

[0099] in:

[0100] X is the original data value,

[0101] Xmin is the minimum value in the data set.

[0102] Xmax is the maximum value in the data set,

[0103] Xnorm is the normalized data value.

[0104] It should be noted that the choice of normalization method should be based on data characteristics and the needs of subsequent analysis. Different normalization methods (such as minimum-maximum normalization, Z-score normalization, etc.) are suitable for different scenarios.

[0105] The purpose of normalization is to simplify calculations and make data processing more convenient and faster. For example, data can be mapped to a range of 0 to 1 for processing, or dimensioned expressions can be converted to dimensionless expressions. In statistics, normalization is used to summarize the statistical distribution of a unified sample. Before normalization, check whether the data has missing values ​​or outliers, which may require data cleaning first.

[0106] Sample element runs test: The runs test is a non-parametric test method used to test the randomness of data. The most direct function of the runs test is to test the randomness of the data sequence. If the sequence is random, then the number and length of the runs should conform to the expected distribution of the random sequence. Through the runs test, it is possible to identify whether there are non-random patterns or trends in the data sequence. For example, if the runs of the geochemical element values ​​are particularly long or short, this may indicate that there is some systematic deviation in the geochemical numbers. The purpose is to select a group of geochemical numbers with significant randomness of geochemical elements through the runs test.

[0107] The advantage of this test is that it does not depend on the distribution form of the data, so it is particularly suitable for data sets that do not satisfy the normal distribution. The basic idea of ​​the runs test is to check whether the number of consecutive occurrences of the same feature in a sequence (run) is randomly distributed. This test method is used to determine the randomness of geochemical elements, which is convenient for carrying out what technical methods to process and analyze data. The calculation formula for the runs test is as follows:

[0108] R = total number of runs;

[0109] n1 = number of observations in the first category;

[0110] n2 = number of observations in the second category;

[0111] The expected number of runs E(R) and variance V(R) are:

[0112]

[0113]

[0114] Among them, n1+n2=n is the total sample size.

[0115] The runs test can be used to test whether a series of data is randomly distributed. If the data is random, the number of runs should be close to the expected number of runs. It should be noted that the runs test does not have strict requirements on the distribution of data, but it has weaker explanatory power for data sets with smaller sample sizes. Therefore, when using the runs test, the results should be interpreted with caution and combined with other statistical methods and professional field knowledge.

[0116] The sample element descriptive statistical analysis is to perform statistics on the formed data. Descriptive statistical analysis is a basic step in data analysis. It provides a preliminary understanding of the overall geochemical data by summarizing and describing the basic characteristics of the data. Through descriptive statistics, errors, outliers or unreasonable records in the data can be found, so as to clean the data. Descriptive statistics can provide preliminary evidence and support for subsequent hypothesis testing. The main functions of this analysis are:

[0117] An overview of the overall geodata can provide basic information about the data, such as the central trend (mean, median, mode) and dispersion (standard deviation, variance, range) of the data.

[0118] The degree of linear correlation between two variables is intended to test whether there is a correlation between the variables. The Pearson correlation coefficient is used to statistically analyze the elements in different periods. There is a correlation between different elements and the same elements in different periods. The changes in elements in the same plot location at different periods and the amplitude characteristics between different elements. The more significant the amplitude characteristics between elements with a smaller change rate, the more favorable it is for the element traceability factor. Ensure that the independent variable elements are correlated with the second-period elements and the dependent variable, and at the same time, the independent variables are independent of each other. Elements with good correlation can reflect homology and have a certain correlation in cause.

[0119]

[0120] Note: ** indicates significant correlation at the 0.05 level (two-tailed), * indicates significant correlation at the 0.01 level (two-tailed).

[0121] Display the shape of the data distribution, including skewness and kurtosis, and visually display the data distribution through histograms, box plots, etc. Use statistics to help identify outliers or extreme values ​​in the data, which may have a significant impact on subsequent analysis, and use cross-tabulations, correlation analysis, etc. to preliminarily explore the relationship between variables.

[0122] Kendall's rank correlation coefficient test is a nonparametric statistical test method used to evaluate whether two random variables are statistically correlated. It is a test method for the rank correlation coefficient. The following are the basic calculation steps of the Kendall τ coefficient based on two variables X and Y:

[0123]

[0124] Where C is the number of concordant pairs, D is the number of dissimilar pairs, and n is the number of data points. A significance test was performed, and the z-score of τ was calculated and compared to a standard normal distribution to determine the significance of the correlation.

[0125] Main functions:

[0126] 1. Evaluate correlation. The Kendall test is used to evaluate whether there is a statistical correlation between two variables, especially when the data does not meet the prerequisites for parametric tests.

[0127] 2. Outlier robustness: Since it is a non-parametric test, the Kendall test is relatively less affected by outliers, which makes it more reliable when outliers are present in the data.

[0128] 3. Small sample applicability: When the sample size is small, the Kendall test can still provide valid statistical results.

[0129] A coefficient greater than 0.05 indicates that there is no significant difference in the data. The significance of this element result is 0.0001. A coefficient less than 0.05 proves that the data has fluctuation differences and inconsistent results, so the next step is to participate in the second-order cluster analysis.

[0130] The paired sample T test is used to compare whether there is a significant difference in the means of two related sample groups, and to detect whether there is a statistically significant difference in the means of two related samples (for example, the measurement results of the same group of subjects under two different conditions). It is suitable for paired design research, such as pre-test and post-test design, repeated measurement design or paired comparison design. It can be used to perform hypothesis testing to determine whether the analysis results of multiple periods of geochemical samples have non-significant correlation characteristics, and whether there is a strong correlation between geochemical elements in the same period.

[0131] Data preparation, ensuring that there are two sets of data, each consisting of the same number of observations, which are paired, such as measurements of the same group of subjects under different conditions.

[0132] The test premise is that the data differences of each pair should be independent and come from a normal distribution. The overall variances of the paired differences should be equal (homogeneity of variance).

[0133] Calculation steps:

[0134] F1: Calculate paired differences: For each pair of observations, calculate their difference;

[0135] F2: Calculate the mean of the differences: Calculate the average of all paired differences;

[0136] F3: Calculate the standard deviation of the difference: Calculate the standard deviation of all paired differences;

[0137] F4: The formula for calculating the T statistic is as follows:

[0138]

[0139] in:

[0140] is the sample mean of the differences;

[0141] μ0 is the population mean of the differences;

[0142] SD is the standard deviation of the difference;

[0143] n is the number of pairs;

[0144] F5: Determine the degrees of freedom: The degrees of freedom (df) is the number of pairs minus 1(n-1).

[0145] If the p-value is less than the pre-set significance level (for example, 0.05), the null hypothesis is rejected and it is considered that there is a significant difference in the means of the two sample groups. If the p-value is greater than the significance level, the null hypothesis cannot be rejected and it is considered that there is insufficient evidence to show that there is a significant difference in the means of the two sample groups.

[0146] The method for testing the correlation between a plot and an element comprises:

[0147] B1: Data collection, collect two sets of data, the two sets of data must be related, such as the measurement results of the same group of subjects under different conditions;

[0148] B2: Calculate paired differences. For each pair of data, calculate their difference. For example, if you compare the test geochemical elements of the same group of subjects before and after treatment, the difference is the first period element of the post-test minus the second period element;

[0149] B3: Check the distribution of the differences to see if they are normally distributed. If the sample size is large enough (usually n>30), the differences can be considered to be approximately normally distributed.

[0150] B4: Calculate statistics, including the mean value Dˉ and standard deviation SD of the difference. The main formula is:

[0151]

[0152] Where n is the number of paired samples;

[0153] B5: Determine the degrees of freedom, the degrees of freedom (df) is the number of paired samples minus 1(n-1);

[0154] B6: Conduct hypothesis test, set the null hypothesis (H0): there is no significant difference between the mean values ​​of the two samples, set the alternative hypothesis (H1): there is a significant difference between the mean values ​​of the two samples, and find the corresponding p value in the T distribution table based on the T statistic and degrees of freedom;

[0155] B7: Draw a conclusion: If the p-value is less than the significance level, reject the null hypothesis and conclude that there is a significant difference between the mean values ​​of the two samples. If the p-value is greater than the significance level, fail to reject the null hypothesis and conclude that there is insufficient evidence to show that there is a significant difference between the mean values ​​of the two samples.

[0156] The sample element second-order cluster analysis includes a pre-clustering stage and a model clustering stage.

[0157] In the pre-clustering stage, a distance-based algorithm is used to create smaller clusters in order to reduce the amount of subsequent calculations. In the model clustering stage, the intra-class covariance matrix is ​​used. The intra-class covariance matrix of each cluster is used to describe the variability within the cluster. The element clustering uses the model selection criteria to determine the final number of clusters and cluster members. The Bayesian Information Criterion (BIC) is used to evaluate the goodness of fit of different clustering models. The BIC formula is as follows:

[0158] BIC=-2log(L)+klog(n)

[0159] in,

[0160] L is the maximum likelihood estimate of the model,

[0161] k is the number of free parameters in the model,

[0162] n is the sample size.

[0163] When calculating BIC, a distance metric is used to determine cluster membership, which is Euclidean distance, Manhattan distance, etc. Identify more detailed element contribution coefficient subgroups, or optimize the original clustering results to identify element combinations with higher similarity within the element variables and more obvious differences between groups.

[0164] The construction of the traceability model includes second-order cluster analysis of sample elements, identification of more detailed element contribution coefficient subgroups, or optimization of the original clustering results. Through second-order cluster analysis, element combinations with higher similarity within element variables Cd, Cr, Pb, Cu, Ni, Se, F, and B and more obvious differences between groups are identified. The sample horizontal positioning inspection model and the sample element vertical plot inspection model are established.

[0165] According to the element contribution ratio, the ratio of cluster size is divided into 2.0%, 10.0%, 42.0%, and 46.0%. Cluster number 1 is the best, cluster similarity of clusters 2 and 3 is in the middle, and cluster 4 has the weakest homology. According to the traceability characteristics of geochemical elements, clusters that are independent of each other are the fourth category, and it is more appropriate to select 2.0% element positioning combination for traceability.

[0166] The Wilcoxon signed Rank Test is a non-parametric statistical test method used to test whether there is a significant difference in the number of locations between two related samples, matched samples, repeated measurement samples or paired samples. The null hypothesis of the hypothesis test is that the two groups of samples come from the same population, that is, the number of locations of the two groups of samples is equal. The alternative hypothesis is that the two groups of samples come from different populations, that is, the number of locations of the two groups of samples is not equal. The result of the positioning element test in the overall plot element shows the constraint range of the positioning element in the vertical value of the overall element as the result of the horizontal comparative analysis.

[0167] The method for constructing the sample lateral positioning inspection model comprises:

[0168] C1: Data collection, collect two sets of data, each set contains the same number of observations, these observations are paired, such as the measurement results of the same group of subjects under different conditions;

[0169] C2: Calculate the paired difference. For each pair of observations, calculate their difference.

[0170] C3: Calculate the sign rank and absolute rank of the difference. For each difference, calculate its sign rank (positive rank, negative rank or zero), as well as the absolute rank;

[0171] C4: Calculate statistics, calculate the rank sum (RS), including the positive rank sum (R+) and the negative rank sum (R-). Use the following formula to calculate the T statistic:

[0172]

[0173] Where n is the number of pairs;

[0174] C5: Determine the degrees of freedom, the degrees of freedom (df) is the number of pairs minus 1(n-1);

[0175] C6: Conduct hypothesis test. The null hypothesis (H0) is that there is no significant difference in the mean values ​​of the two samples. The alternative hypothesis (H1) is that there is a significant difference in the mean values ​​of the two samples. Based on the T statistic and degrees of freedom, find the corresponding p-value in the Wilcoxon distribution table.

[0176] C7: Draw a conclusion. If the p-value is less than the significance level, reject the null hypothesis and conclude that there is a significant difference between the mean values ​​of the two samples. If the p-value is greater than the significance level, fail to reject the null hypothesis and conclude that there is insufficient evidence to show that there is a significant difference between the mean values ​​of the two samples.

[0177] C8: Conduct geochemical analysis based on the detection characteristics;

[0178] The sample element longitudinal plot test model is established to test whether there is a significant difference in the medians of three or more independent samples.

[0179]

[0180] in,

[0181] N is the total number of observations,

[0182] k is the number of samples,

[0183] ni is the number of observations in the ith sample,

[0184] Ri is the rank sum of the i-th sample,

[0185] According to the value of the H statistic and the corresponding distribution table, determine whether to reject the null hypothesis, that is, the number of locations of all samples is equal. If the p value is less than the pre-set significance level, the null hypothesis is rejected, and it is considered that the number of locations of at least one sample is different from that of other samples.

[0186] The Kruskal-Wallis null hypothesis is that all samples come from the same population, while the alternative hypothesis is that at least one sample comes from a different population. This test is based on comparing the rank values ​​of all samples, rather than comparing the original data. This method can be used to test whether the elements come from the same population, or whether the median of a population is different from the location number of other populations. The use of non-parametric verification of geochemical numbers is mainly used to judge the status of elements in the overall sample range based on the comparison of multiple periods of geochemical element samples, correlation analysis, and test distribution differences.

[0187] After the traceability model is established, it is tested. According to the above technical processing and analysis, the geochemical element traceability positioning code macro command method is used to locate the specific plot location. Through step-by-step comparison by computer, the specific element value can be determined.

[0188] When tracing the soil, the Wilcoxon sign test method is used to locate the elements of the overall plot element detection constraint range and the Kruskal-Wallis test element plot method, presenting the positioning elements to constrain the value of the plot element correlation and distance detection results. The value selection principle selects the element values ​​on both sides of the blue line and the green line, and it is appropriate to select the value of =>0.5 on both sides of the line. According to the value selection principle, the second-phase positioning elements are the positioning number 1.0464 on both sides of the B element value range of -1 to 1.5, the positioning number -0.5688 in the Cd element value range of -0.5 to -1.0 , the positioning number 1.1398 is in the range of -0.5 to 1.5 for the Cr element, the positioning number 0.5892 is in the range of -0.6 to 1.0 for the Cu element, the positioning number 1.0940 is in the range of -0.5 to 1.6 for the F element, and the positioning number 1.2110 is in the range of -0.1 to 1.7 for the Ni element. According to the comparative analysis of 92 elements, the positioning numbers are in line with the first-phase elements of 31, 49, 31, 38, 41, and 34. The values ​​of the elements that meet the requirements are used as the basic data for plot distribution display, such as Figure 6 As shown;

[0189] Through non-parametric test methods, the elements are distributed in the range of 10 plots in the element test method. Multiple types of elements have element information values. Through step-by-step comparison by VBA macro code, the specific element values ​​can be determined. The LoadTrainData and LoadTrainLabels functions of the 1.1 positioning method are mainly implemented to load the corresponding element data, training labels, test data and test labels. The positioning elements are gradually significant in the E, C, and M class plots among the 15 plots. Combining the element and plot positioning method, the positioning numbers are judged to be 1.0464, -0.5688, 1.1398, 0.5892, 1.0940, and 1.2110. This group of geochemical elements is strongly significant for the E1 class 11 collection plot. Figure 7 shown.

[0190] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0191] In the description of the present invention, the description with reference to the terms "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0192] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

Claims

1. A method for tracing the origin of elemental soil plots based on non-parametric testing, characterized in that: The following steps are involved: A1: Carry out the first soil collection and test the geochemical elements in the first soil collection sample; A2: Conduct a second soil collection and test the geochemical elements in the second soil collection sample at the same location; A3: Process the data of the first phase and the second phase; A4: Compare and analyze the data of Phase I and Phase II and build a traceability model; A5: Test the geochemical elements of the soil samples that need to be traced, bring the test data into the traceability model for statistical analysis, and determine the location of the element numbers belonging to the collection site of the soil sample; The analysis and processing of the first-phase data and the second-phase data include data processing by sample data normalization method, sample element runs test, sample element descriptive statistical analysis, sample element Kendall test, first-phase and second-phase sample element paired sample correlation analysis, T test, and plot and element correlation test.

2. The element soil plot tracing method based on non-parametric test according to claim 1 is characterized in that: The formula for data processing of the sample data normalization method is as follows: in: X is the original data value, Xmin is the minimum value in the data set. Xmax is the maximum value in the data set, Xnorm is the normalized data value.

3. The element soil plot tracing method based on non-parametric test according to claim 2 is characterized in that: The calculation formula for the sample element runs test is as follows: R = total number of runs; n1 = number of observations in the first category; n2 = number of observations in the second category; The expected number of runs E(R) and variance V(R) are: Among them, n1+n2=n is the total sample size.

4. The element soil plot tracing method based on non-parametric test according to claim 3 is characterized in that: The sample element descriptive statistical analysis performs statistics on the generated data, and provides a preliminary understanding of the overall geochemical data by summarizing and describing the basic characteristics of the data. Through descriptive statistics, errors, outliers or unreasonable records in the data can be discovered, thereby performing data cleaning. Descriptive statistics can provide preliminary evidence and support for subsequent hypothesis testing.

5. The element soil plot tracing method based on non-parametric test according to claim 4 is characterized in that: The sample element Kendall test is used to evaluate whether two random variables are statistically consistent. The following are the steps for calculating the Kendall W coefficient based on two variables X and Y: Where n is the number of subjects, k is the number of raters, is the average rating of the ith object, It is the average of the average scores of all objects. The value range of Kendall's W coefficient is between 0 and 1. The closer the value is to 1, the higher the consistency between the scorers, and the closer the value is to 0, the lower the consistency. It is compared with the standard normal distribution to determine the fluctuation difference of element data and prove the significant characteristics of inconsistency, which can provide data support for subsequent cluster analysis.

6. The element soil plot tracing method based on non-parametric test according to claim 5 is characterized in that: The paired sample T test of the first and second phase sample elements comprises the following steps: F1: Calculate paired differences: For each pair of observations, calculate their difference; F2: Calculate the mean of the differences: Calculate the average of all paired differences; F3: Calculate the standard deviation of the difference: Calculate the standard deviation of all paired differences; F4: The formula for calculating the T statistic is as follows: in: is the sample mean of the differences; μ0 is the population mean of the differences; SD is the standard deviation of the difference; n is the number of pairs; F5: Determine the degrees of freedom: The degrees of freedom (df) is the number of pairs minus 1(n-1).

7. The element soil plot tracing method based on non-parametric test according to claim 6 is characterized in that: The method for testing the correlation between a plot and an element comprises: B1: Data collection, collect two sets of data; B2: Calculate the paired difference. For each pair of data, calculate their difference; B3: Test the distribution of differences and whether these differences obey normal distribution; B4: Calculate statistics, including the mean value Dˉ and standard deviation SD of the difference. The main formula is: Where n is the number of paired samples; B5: Determine the degrees of freedom, the degrees of freedom (df) is the number of paired samples minus 1(n-1); B6: Conduct hypothesis test, set the null hypothesis (H0): there is no significant difference between the mean values ​​of the two samples, set the alternative hypothesis (H1): there is a significant difference between the mean values ​​of the two samples, and find the corresponding p value in the T distribution table based on the T statistic and degrees of freedom; B7: Draw a conclusion: If the p-value is less than the significance level, reject the null hypothesis and conclude that there is a significant difference between the mean values ​​of the two samples. If the p-value is greater than the significance level, fail to reject the null hypothesis and conclude that there is insufficient evidence to show that there is a significant difference between the mean values ​​of the two samples.

8. The method for tracing the origin of elemental soil plots based on non-parametric testing according to claim 7 is characterized in that: The sample element second-order cluster analysis includes a pre-clustering stage and a model clustering stage. In the pre-clustering stage, a distance-based algorithm is used to create smaller clusters in order to reduce the amount of subsequent calculations. In the model clustering stage, the intra-class covariance matrix is ​​used. The intra-class covariance matrix of each cluster is used to describe the variability within the cluster. The element clustering uses the model selection criteria to determine the final number of clusters and cluster members. The Bayesian Information Criterion (BIC) is used to evaluate the goodness of fit of different clustering models. The BIC formula is as follows: BIC=-2log(L)+klog(n) in, L is the maximum likelihood estimate of the model, k is the number of free parameters in the model, n is the sample size.

9. The element soil plot tracing method based on non-parametric test according to claim 8 is characterized in that: The method for constructing the sample lateral positioning inspection model comprises: C1: Data collection, collect two sets of data, each set of data contains the same number of observations, and these observations are paired; C2: Calculate the paired difference. For each pair of observations, calculate their difference. C3: Calculate the sign rank and absolute rank of the difference. For each difference, calculate its sign rank and absolute rank. C4: Calculate statistics, calculate the rank sum (RS), including the positive rank sum (R+) and the negative rank sum (R-). Use the following formula to calculate the T statistic: Where n is the number of pairs; C5: Determine the degrees of freedom, the degrees of freedom (df) is the number of pairs minus 1(n-1); C6: Conduct hypothesis test, null hypothesis (H0): there is no significant difference in the mean values ​​of the two samples, alternative hypothesis (H1): there is a significant difference in the mean values ​​of the two samples, find the corresponding p value in the Wilcoxon distribution table based on the T statistic and degrees of freedom; C7: Draw a conclusion. If the p-value is less than the significance level, reject the null hypothesis and conclude that there is a significant difference between the mean values ​​of the two samples. If the p-value is greater than the significance level, fail to reject the null hypothesis and conclude that there is insufficient evidence to show that there is a significant difference between the mean values ​​of the two samples. C8: Perform geochemical analysis based on the detection characteristics.

10. The method for tracing the origin of elemental soil plots based on non-parametric testing according to claim 9, characterized in that: The sample element longitudinal plot test model is established to test whether there is a significant difference in the medians of three or more independent samples. in, N is the total number of observations, k is the number of samples, ni is the number of observations in the ith sample, Ri is the rank sum of the i-th sample, According to the value of the H statistic and the corresponding distribution table, determine whether to reject the null hypothesis, that is, the number of locations of all samples is equal. If the p value is less than the pre-set significance level, the null hypothesis is rejected, and it is considered that the number of locations of at least one sample is different from that of other samples.

Images